On September 29, 2026, OpenAI officially changed its ChatGPT Pro lineup. Pro 500 was introduced at $500/month, and Pro 200 became available again at $200/month with changed included usage allowances. OpenAI’s own Help Center states that new Pro 200 subscriptions that are not grandfathered receive a lower included usage allowance than the previous Pro 200 plan, while the monthly price remains $200.
I am one of the affected $200 Pro users.
My Pro 200 upgrade / weekly usage reset happened at 6:09 PM on September 29.
I then used ChatGPT Work mainly from about 6:09 PM until about 9:09 PM — roughly three hours total — for eight chapters of writing and review work.
At approximately 9:09 PM, I stopped using Work.
The screenshot showing 89% of my weekly allowance remaining was taken at around 10:30 PM while I was preparing and submitting feedback about the issue.
To be absolutely clear: 10:30 PM was only the time the evidence screenshot was taken while submitting feedback. It was NOT the time I stopped using Work. My actual Work usage had already ended at approximately 9:09 PM, more than an hour earlier.
The timeline is therefore:
• 6:09 PM — Pro 200 upgrade / weekly reset
• About 6:09 PM–9:09 PM — approximately three hours of actual Work usage
• About 9:09 PM — I stopped using Work
• Around 10:30 PM — while submitting feedback, I took the screenshot showing only 89% of the weekly allowance remaining
This is not just one isolated complaint.
As of September 29, our public-source tally has already identified and de-duplicated at least 45 affected users reporting similar concerns across $100, $200, and $500 Pro plans — including unusually fast quota depletion, usage/accounting inconsistencies, paid-plan allowances that appear not to match expectations, and usage limits that users cannot clearly reconcile with their actual activity.
That 45-user figure is not an official OpenAI number and is not meant to represent every affected user worldwide. It is simply the number of distinct publicly identifiable cases we have been able to verify so far, and the number is still increasing.
I am also not claiming that every one of these reports has the exact same technical root cause.
What I am asking OpenAI to do is simple:
- Clearly explain how ChatGPT Work and Codex usage is calculated.
- Show task-level or session-level usage deductions.
- Show which model, reasoning level, retries, background activity, subagents, compactions, or other operations consumed the allowance.
- Clearly explain how the new $100, $200, and $500 Pro allowances compare.
- Allow paying customers to audit exactly where their weekly usage went.
- Investigate cases where the visible usage appears inconsistent with the work the user actually performed.
When customers are paying $100, $200, or $500 every month, usage accounting should be transparent, understandable, and auditable.
OpenAI should provide a clear public explanation.